Architecture Overview for Data Analytics Deployments Mahmoud Ghanem Sr. Systems Engineer GLOBAL SPONSORS
Agenda The Big Picture Top Use Cases for Data Analytics Modern Architecture Concepts for Data Analytics Dell EMC Solutions to Simplify Data Analytics 2
The Big Picture DATA ANALYTICS DEVOPS CLOUD INDUSTRY LEADER + + = KNOW SOONER ACT FASTER SCALE IT DIGITAL DISRUPTION 3
DATA ANALYTICS SPECTRUM DATA MINING MACHINE LEARNING GENERAL AI DESCRIPTIVE STATISTICS PREDICTIVE MODELING PRESCRIPTIVE RECOMMENDATIONS GENERAL INTELLIGENCE 4
Top Use Cases for Data Analytics EDW Optimization 360 View of Customer Security, Risk & Compliance ML IOT Predictive Analytics 5 Operations Intelligence
DATA ANALYTICS PLATFORM SECURITY APPLICATION LAYER Dashboards Visualization Platform As A Service Data Marketplace Application Marketplace Collaboration Real time Control DATA LAYER Data Discovery Search Analytics (Graph, Machine Learning, Spatial, Sentiment) API Management (Data API, Micro-Services) Catalog ENABLE DATA FOR CONSUMPTION ANALYTICS / DATA APIS Data Ingestion (IoT, Social, City, Personal) Data Stores (Structured, Unstructured, Object, In-Memory, Key Value) Data Wrangling (Transformation, Cleansing) Data Governance (Audit, Workflow, Encryption) OPEN STANDARDS DATA INGESTION, STORE, CLEANSE & GOVERNANCE INFRASTRUCTURE LAYER INFRASTRUCTURE FOUNDATION Compute (Converged, Scale-up, Scale-out) Network (SDN,SAN, WAN, Wireless) Storage (SDS, Block, File, Object) Cloud M&O (Private, Public) Business Continuity (Backup & Recovery, DR)
A strategic comparison of modern architecture concepts for Data Analytics. 7
BUY BUILD 8
BUY 9
BUY Time to value Commodity use case Simplicity at a cost Incumbent evolution/expansion Existing talent Use/source diversity is low 10
11 BUILD
12 BUILD Unique value stream Snowflake use case Cheap = Complex Lower cost of incumbents Talent and DevOps culture Massive scale and variety
BATCH STREAMING 13
BATCH 14
BATCH Descriptive Fidelity matters Large volumes Data Science playground Time is relative Scheduled 15
STREAMING 16
Potentially Predictive Speed trumps fidelity Parallel for streams Data Science outcomes Talent and DevOps culture Massive scale and variety STREAMING 17
PUBLIC PRIVATE 18
VIRTUAL PHYSICAL 19
DAS NAS 20
21
The Ready Solutions formula Dell EMC portfolio Priorities Compute Deploy Ready Nodes Ready Bundles Ready Systems Biz Apps Knowledge Services 22
Dell EMC Ready Bundles for Hadoop Dell EMC Ready Bundle for Cloudera Hadoop End-to-end data management, processing and analytics with no-code-needed deployment plus high double-digit performance gains Dell EMC Ready Bundle for Hortonworks with Isilon Capacity optimized and efficient data processing and data lake platform leveraging Isilon for shared storage for Hadoop with HDFS as a protocol. Dell EMC Ready Bundle for Hortonworks Hadoop Open Source data management, processing and analytics solution that efficiently process multistructured data volumes using existing tools and resources 23
Big Data Technology Advisory Service Develop an architecture and plan to implement big data capabilities Analytics Application development Statistical Modeling/Natural Language Processing/ Machine Learning Enterprise Search/Index Data Exploration/ Visualization Data Warehousing Data Discovery Data Transformation Business Intelligence Data Tagging / Metadata Management Hadoop / SQL on Hadoop Data Ingestion Enterprise Log Analysis Key Steps: 1. Determine Target Capabilities and Outcomes 2. Assess Current State 3. Determine Future State Architecture 4. Perform Gap Analysis 5. Recommend Architecture Roadmap and plan 24
The journey is worth it 25